As already mentioned, many species that have wide breeding ranges show a 'migratory divide'. On one side of the divide birds head in one direction to winter quarters, while on the other side they head in a different direction. Often (but not always), such divides lead birds to circumvent geographical obstacles such as seas, deserts or mountain ranges. They have been described from ringing recoveries in many European species that fly to Africa via the shortest sea-crossings at the western and eastern ends of the Mediterranean, including the White Stork Ciconia ciconia, Black Stork C. nigra, various soaring raptors, and some small warblers, such as the Blackcap Sylvia atricapilla and Greater Whitethoat Sylvia communis (Figures 8.5 and 22.2). Hence, while one explanation of migratory divides is that they represent the meeting point of populations that spread in post-glacial times from different glacial refuges but have retained their ancestral migration routes, another explanation is that they evolved more recently to circumvent barriers. The two explanations are not mutually exclusive, for while divides may persist as post-glacial legacies, they may now serve a secondary role in directing birds around unfavourable areas. However, the taxonomic or genetic differentiation in many species on either side of the divide implies long-term independent evolution, and hence their refuge-based origin. Once again, their current migration routes follow their likely past colonisation routes. In some species, moreover, DNA analyses have dated the time of the separation of west and east populations roughly to the last glacial period when they were refuge-based (as for Swainson's Thrush Catharus ustulatus discussed above).
In North America, the explanation of migratory divides is less equivocal because no major ecological barrier lies west-east across the main flyways north of the Gulf of Mexico. Yet many birds were confined to southwest and southeast parts of the continent during glacial times, and still have distinct southwestern and southeastern wintering areas. Compared with the species mentioned earlier, the Tundra Swan Cygnus columbianus is an extreme example, breeding across much of the North American tundra, but wintering on either the west or east coasts. The migratory divide lies far to the west, at Point Hope in Alaska, and the bulk of the population (including that part directly north of the Californian wintering area) winters on the east coast (Figure 22.7).
Whatever the origin of migratory divides, an interesting question is how they persist, given that directional preferences are inherited, and the populations on either side can (at least in theory) interbreed where they meet. The persistence of the divide may be due to the fact that migratory directions and distances inherited by hybrids are intermediate between those of their parents (Chapter 20). Such hybrids
Figure 22.7 Breeding and wintering ranges of the Tundra Swans Cygnus columbianus in North America, showing the migratory divide in Alaska. Birds from the southwest of the breeding range in Alaska winter on the western side of the continent, but birds from most of the breeding range winter on the eastern side of the continent.
arising in the contact zone may then be less viable than the parental forms, contributing to the maintenance of the divide, and the genetic integrity of both populations. In this case, migratory divides might contribute to the speciation process through reducing interbreeding between the populations on either side of the divide.
One of the best documented examples of non-hybridisation concerns the Greenish Warbler Phylloscopus trochiloides, in which populations spreading northwards in post-glacial times around the west and east sides of the Tibetan Plateau met on the north side of the Plateau, in southern Siberia (Irwin et al. 2001). During their isolation, the two forms had not only developed minor genetic differences (recognisable in the mitochondrial DNA), but also different songs. Where they meet on the north side, birds from the two populations do not recognise one another's song, and so remain distinct without any significant interbreeding. In effect, they behave where they meet as different species, but are linked by a ring of interbreeding populations encircling the south side of the Plateau. The two forms are assumed to migrate southward down the west and east sides of the Plateau to reach their wintering areas. The Tibetan Plateau is evidently a major barrier to migration. Of 97 longdistance migrants breeding in Siberia, most (85%) use only one route round Tibet (42 through Kazakhstan, 40 through eastern China). Of the 15 species that use both routes, seven are known to have migratory divides between western and eastern subspecies (Irwin & Irwin 2005). In four additional cases, migratory divides occur between western and eastern sister species. These findings suggest that two very different migratory programmes seldom persist in the same gene pool, and that migration may play a role in speciation in this region. Moreover, although all described divides in migratory birds relate to breeding areas, equivalent divides must presumably operate in some wintering areas wherever adjacent populations separate to reach their individual breeding areas.
Sometimes bird populations nesting relatively close to one another have totally different migration patterns, marked by a divide. For example, Harlequin Ducks Histrionicus histrionicus nesting on the east side of Labrador migrate south to winter on the coast of Maine, whereas those from northern Quebec and Labrador migrate northeast to winter on the coast of Greenland (being one of the few bird populations that winters at higher latitudes than its breeding area (Brodeur et al. 2002). The authors suggested that these two population segments may represent previously isolated populations, one originating from the Pleistocene glacial refuge in western Greenland and the other from south of the Laurentide ice sheet in eastern North America. It is not known whether other species share this divide, or whether it is confined to Harlequin Ducks.
Other migratory divides are evident in the far north of Eurasia and North America, especially among seabirds and waders which, on leaving the tundra, initially fly along the northern coasts before turning south down the Atlantic or Pacific coastlines. In Eurasia, many species show an abrupt migratory divide at about 100°E on the Taimyr Peninsula, with post-breeding movement west of that site being mainly west-southwest and east of that site mainly east-southeast, more or less parallel to the coast (Alerstam & Gudmundsson 1999). One plausible explanation for the divide being situated in the Taimyr Peninsula is that this region lies midway across the Eurasian land mass. Hence, in whatever way this divide evolved, it may persist largely as a result of distance-dependent flight costs
(Alerstam & Gudmundsson 1999). However, not all tundra species show a divide at this point, and some take more southerly routes across the Eurasian land mass to their winter quarters. Any shorebirds migrating from Taimyr to India, say, would have to cross at least 5000 km of the Eurasian land mass before reaching another marine coastal site. Other migratory divides occur in Greenland, with birds from west and southern parts of the island wintering in North America, and those from the northeast wintering in Europe (for Snow Buntings Plectrophenax nivalis and others see Lyngs 2003).
Even in the absence of geographical barriers, or twin glacial refuges, it is not hard to see how migratory divides persist. Imagine a species with a wide breeding range that has two wintering areas near either end, as in Figure 22.8. Imagine now that migration costs increase with length of journey, in proportion to the lengths of lines in Figure 22.8a. If birds with higher migration costs return in smaller proportion to the breeding range, generation after generation, then segregation of breeding populations with a clear migratory divide will develop (Figure 22.8b). If competition on wintering areas is intense, this could lead to a gap in the breeding range of the species, because individuals migrating the longest distances to fill that gap could be disadvantaged during competition on the wintering range, and continually be eliminated by selection.
What could change this situation, leading to breakdown of the divide, is if the conditions in one wintering area were so good that birds from all parts of the breeding range benefited from migrating there. A migratory divide could shift by populations changing their inherited migration direction under the action of natural selection, but a more likely mechanism would be through the spread of the more successful wintering population into the breeding range of the other, gradually replacing it (Figure 22.8c).
Another type of divide exists in populations in which some individuals take one direction to a wintering area, while others from the same place take a different direction to another wintering area. An example is provided by the central European Blackcaps Sylvia atricapilla, some of which began to migrate westward to winter in Britain in the mid-twentieth century (Chapter 20). These birds differ from those that continue to migrate southwest to Iberia. Although birds from both types currently breed in the same general area, they show assortative mating, pairing with birds of their own directional preference much more than expected by chance, owing to a difference in arrival times (Bearhop et al. 2005). The divide may be temporary, and in time one migratory genotype may gradually replace the other.
In conclusion, the various migratory routes and patterns of distribution described above could all represent a carry-over from conditions operating at an earlier stage in the earth's history. They would be unlikely to persist, however, if they were markedly disadvantageous in present conditions. Nonetheless, the reason that some birds stick to their present migration routes may stem from the difficulty of making a single large-step change in migratory habits that would otherwise be beneficial: for example, a big change in direction that would greatly reduce the length of the migratory journey. For most of the patterns described above, alternative explanations have been proposed. They again involve untested assumptions and, although sometimes couched in mathematical terms, they are still little more than guesses. However, the idea that glacial conditions and post-glacial colonisation
Figure 22.8 Model depicting the evolution and maintenance of a migratory divide without the necessary involvement of glacial refuge areas. (a) Hypothetical starting scenario, with a continuous breeding range, and two separate wintering areas, each visited by birds from all parts of the breeding range. Migration costs proportional to lengths of lines depicting routes. (b) Migratory divide resulting from differential survival consequent upon differential migration costs (or from migration costs plus asymmetrical competition favouring use of nearest breeding and wintering sites). In this model the carrying capacity of the two wintering areas is assumed to be equal. The divide then appears midway through the breeding range, so that the populations on each side are of about equal size. (c) Situation resulting from differential survival consequent upon differential migration costs, together with wintering areas of markedly different qualities or carrying capacities. Since the carrying capacities of the two wintering areas are unequal, the migratory divide would be expected to shift from the centre of the breeding range towards whatever side had the wintering area of smallest capacity. Other modifications to the model could be envisaged by changing the relative distances of the wintering sites from the breeding range, and hence the migration costs. Modified from Lundberg & Alerstam (1986).
patterns could have influenced migratory behaviour is increasingly supported by palaeontological and DNA studies.
Was this article helpful?